The present invention relates to an improved process for the production of peptides by solid-phase synthesis. The invention also relates to agents which are useful in solid-phase peptide synthesis.
Solid-phase peptide synthesis (SPPS) is a highly successful method introduced by Merrifield in 1963 (Merrifield, R. B. (1963) J. Amer. Chem. Soc. 85, 2149-2154). Numerous peptides have been synthesized with this technique since then. Methods used in the prior art to chemically synthesize peptides and proteins are reviewed in Kent, S. B. H. (1988), Ann. Rev. Biochem. 57, 957-989.
Two strategies for the assembly of peptide chains by solid-phase synthesis have been used, viz. the stepwise solid-phase synthesis, and solid-phase fragment condensation. In stepwise SPPS, the C-terminal amino acid in the form of an N-α-protected, if necessary, side-chain, protected reactive derivative is covalently coupled either directly or by means of a suitable linker to a “solid”-support, e.g., a polymeric resin, which is swollen in an organic solvent. The N-α-protective group is removed, and the subsequent protected amino acids are added in a stepwise fashion. When the desired peptide chain length has been obtained, the side-chain protective groups are removed, and the peptide is cleaved from the resin. This may be done in separate steps or at the same time. In solid-phase fragment condensation, the target sequence is assembled by consecutive condensation of fragments on a solid support using protected fragments prepared by stepwise SPPS.
Over the years, two coupling strategies have been developed based on the use of different N-α-protective groups and matching side-chain protective groups. Merrifield used tert-butyloxycarbonyl (Boc) as the N-α protective group, while 9-fluorenylmethyloxycarbonyl (Fmoc) was introduced by Carpino and Han (Carpino, L. A. and Hari, G. Y. (1972), J. Org. Chem. 37, 3404-3409). The operations involved in one cycle of chain extension in stepwise SPPC using Boc- and Fmoc-chemistries are illustrated in FIG. 1 (taken from Kent, S. B. H. (1988), Ann. Rev. Biochem. 57, 957-989). The side-chain protection in both cases was tert-butyl, trityl and arylsulfonyl based, and these side chains were deprotected with TFA.
The N-α-Boc-protected peptide coupled to a PAM-resin was N-α-deprotected with trifluoroacetic acid (TFA). The resulting amine salt was washed and neutralized with a tertiary amine. The subsequent peptide bond was formed by reaction with an activated Boc-amino acid, e.g., a symmetric anhydride. Generally, the side-chain protection is benzyl-based, and the deprotection is made with HF or a sulphonic acid.
The N-α-Fmoc protected peptide coupled to a resin was N-α-deprotected by treatment with a secondary amine, normally piperidine, in an organic solvent, e.g., N,N-dimethyl formamide (DMF) or dichloromethane (DCM). After washing, the neutral peptide resin was reacted with an activated Fmoc-amino acid, e.g., a hydroxybenzotriazole active ester.
While the Boc- and Fmoc-strategies have been used for essentially all current practical peptide synthesis, other N-α protective groups have been proposed (Stewart, J. M. and Young, J. D., Solid phase Peptide Synthesis, Pierce Chemical Company (1984)). Boc forms an acid-labile urethane group, and other proposals of this category have been biphenylisopropyloxycarbonyl (Bpoc), 3,5-dimethoxyphenylisopropyloxycarbonyl (Ddz), phenylisopropyloxycarbonyl (Poc) and 2,3,5-tetramethylbenzyloxycarboxyl (Tmz). Other types of N-α protecting groups available include nitrophenylsulfenyl (Nps) which can be removed by either very dilute anhydrous acid, e.g., HCl, or by nucleophilic attack, e.g., with methyl-3-nitro-4-mercapto benzoate. Also dithiasuccinyl (Dts), which is removable by nucleophilic attack, might be used.
SPPS has the general advantage that it lends itself to fully automated or semi-automated chain assembly chemistry. A system for SPPS under low pressure continuous flow conditions was developed by Dryland and Sheppard (1986) J. Chem. Soc. Perkin Trans. I, 125-137 and was further refined (Cameron, L., Meldal, M. and Sheppard, R. C (1987), J. Chem. Soc. Chem. Commun. 270-272; Meldal et al. (1993), Int. J. Pept. Protein Res. 41, 250-260; and WO 90/02605). While SPPS has now developed to be a cornerstone in protein and peptide synthesis, certain problems still remain to be solved. Since some of these problems might well be related to the peptide structure, a brief discussion regarding protein conformation is deemed proper.
Empirical predictions of protein conformations have been made by Chou and Fasman ((1978), Ann. Rev. Biochem. 47, 251-276.). It is well known that protein architectures may be described in terms of primary, secondary, tertiary, and quaternary structure. The primary structure refers to the amino acid sequence of the protein. The secondary structure is the local spatial organization of the polymer backbone without consideration of the side-chain conformation. As examples of secondary structures, α-helices, β-sheets, and β-turns, which are chain reversal regions consisting of tetrapeptides can be mentioned. The tertiary structure is the arrangement of all the atoms in space, including disulphide bridges and side-chain positions, so that all short and long-range interactions are considered. The term quaternary structure may be used to denote the interaction between subunits of the protein, e.g., the α and β-chains of hemoglobin.
Following a discussion of earlier attempts to correlate protein secondary structure with amino acid compositions where, e.g., Ser, Thr, Val, Ile, and Cys were classified as “helix breakers” and Ala, Leu, and Glu as “helix formers,” while hydrophobic residues were classified as strong “β-formers,” and proline together with charged amino acid residues as “β-breakers,” Chou and Fasman made a statistical analysis of 29 proteins with known X-ray structure in order to establish prediction rules for α- and β-regions (Chou and Fasman, 1978, Ann. Rev. Biochem. 47:251-276). Based on these studies, they determined so-called propensity factors Pα, Pβ and Pt which are conformational parameters expressing the positional preferences as α-helix, β-sheet and β-turn, respectively, for the natural L-amino acids forming part of proteins. For the sake of convenience, the Pα and Pβ values are listed below in Table 1. Generally speaking, values below 1.00 indicate that the amino acid in question must be regarded as unfavorable for the particular secondary structure. As an example, the hydrophobic acids (e.g., Val, Ile, Leu) are strong β-sheet formers, while the charged amino acids (e.g., Glu, Asp, His) are β-sheet breakers.
TABLE IPROPENSITY VALUESPαPβGlu1.51Val1.70Met1.45Ile1.60Ala1.42Tyr1.47Leu1.21Phe1.38Lys1.16Trp1.37Phe1.13Leu1.30Gln1.11Cyr1.19Trp1.08Thr1.19Ile1.08Gln1.10Val1.06Met1.05Asp1.01Arg0.93His1.00Asn0.89Arg0.98His0.87Thr0.83Ala0.83Ser0.77Ser0.75Cys0.70Gly0.75Tyr0.69Lys0.74Asn0.67Pro0.55Pro0.57Asp0.54Gly0.57Glu0.37
In the α-helix structure, the spiral configuration of the peptide has been found to be held rigidly in place by hydrogen bonds between the hydrogen atom attached to the nitrogen atom in one repeating unit and the oxygen atom is attached to a carbon atom three units along the chain. If a polypeptide is brought into solution, the α-helix can be made to unwind to form a random coil by adjustment of the pH. The transition from α-helix random coil occurs within a narrow pH. Since the hydrogen bonds are all equivalent in bond strength in the α-helix, they tend to let go all at once. The change can also be induced by heat.
The β-sheet structure consists of fully extended peptide chains in which hydrogen bonds link the hydrogen atoms on one chain to the oxygen atoms in the adjoining chain. Thus hydrogen bonds do not contribute to the internal organization of the chain as they do in the α-helix, but only bond chain to chain. Adjacent chains may be parallel or antiparallel. β-turns are frequently observed in these parts of a peptide chain which connect antiparallel chains in a β-sheet structure. In a β-turn, the CO— and NH-groups from amino acid No. n in the peptide chain form hydrogen bond to the corresponding groups in amino acid No. n+4.
α-Helix and β-sheet constitute strongly varying parts of the peptide conformation of proteins (from 0 to 80%), and the remaining parts of the proteins are folded in other structures. In most proteins, sections of the peptide chains appear as irregularly folded “random coils.”
Turning now to the general problems still prevailing in connection with SPPS, S. B. H. Kent ((1988), Ann. Rev. Biochem. 57, 957-989) highlights the synthesis of “difficult sequences.” Obviously, the whole rationale of SPPS is based on a complete N-α-deprotection prior to each of the coupling steps involved. By the same token, ideally all of the N-α-deprotected amino groups should be coupled to the reactive amino acid derivative according to the desired sequence, i.e., a complete aminoacylation should take place. Kent states that the most serious potential problem in stepwise SPPS is incomplete peptide bond formation giving rise to peptides with one or more amino acids missing (deletions), but with properties similar to the target sequence. Such incomplete couplings are more prevalent in some sequences than in others, hence the term “difficult sequences,” and are apparently also more predominant in Fmoc-chemistry than in Boc-chemistry.
A number of recognized “difficult sequences” have been previously studied. During SPPS of homo-oligopeptides containing leucine or alanine using the Fmoc-strategy, ineffective N-α-deprotection with piperidine in a sequence dependent manner (B. D. Larsen, C. Larsen, and A. Holm in Peptides 1990, E. Giralt and D. Andreu, (Eds). 1991 ESCOM Science Publishers B.V., p. 183-185; and Larsen, B. D., and Holm, A. (1994), Int. J. Pept. Protein Res. 43, 1-9) was observed. Investigations showed that this phenomenon was associated with subsequent slow or incomplete amino acid coupling and evidence for β-sheet aggregation of the growing peptide chain was presented as a cause for the difficult couplings and incomplete Fmoc-deprotections. This evidence was based on general physical-chemical observations (Larsen, B. D., and Holm, A. (1994), Int. J. Peptide & Protein Res. 43, 1-9) and on a detailed Raman Near Infrared Spectroscopic study (Due Larsen, B., Christensen, D-H., Holm, A., Zillmer, R., and Faurskov, O. (1993), J. Am. Chem. Soc. 115, 6247-6253).
Kent ((1988), Ann Rev. Biochem. 57, 957-989) proposes a number of solutions to the problem related to sequence-dependent coupling difficulties, viz. the use of heat in the coupling step and a quantitative conversion of residual unreacted resin-bound peptide chains to terminated species in a “capping” procedure. However, at the present time, no procedure has been formulated for synthesizing these “difficult sequences” in high yield and purity.
It is therefore one object of the present invention to provide an improved SPPS according to which peptides which are recognized as or prove to be “difficult sequences,” can be synthesized in high yield and purity. A further object of the invention is to provide an improved SPPS which provides for reduced coupling times, not only for difficult sequences, but also for otherwise uncomplicated sequences where it is desirable to reduce the normally long coupling times. A still further object of the invention is to provide an agent or a kit for use in SPPS, whereby the above-mentioned process improvements are obtained.